How To Build A Currency Basket

Includes: FXA, FXC, FXF, FXG
by: Alberto Quintana

Imagine our main currency is Euro (but it could be dollars, pounds, the process is exactly the same). We are paid in euros and we live in a Euro country. But we don´t trust our currency that much, or we are not sure where we will live in a few years, or simply to diversify our portfolio, we want to create a currency basket. Problem: which currencies and in which percentage?

There are many tools to get exposure to different currencies, like cash, bonds, etc. Here, we will use ETFs such as FXF for the Swissy, FXA for the Aussie, FXC for the Canadian dollar, cash for the US dollar, and FXG for the pound. In any case, the main point of the article is to show an algorithm to build a reasonable currency basket.

Second, we have picked the currencies based on fundamental reasons and volume, with no emergency countries involved. The result is, as we have seen, Swiss francs, Australian dollar, Canadian dollar, US dollar, and pound. We will return to this selection in another article. (Note: no yen due to awfully bearish perspective.)

To calculate their weights in our basket we will deal with 4 aspects: Purchase Power Parity (PPP), market, inflation, and interest rates.

Using the Big Mac Index as a benchmark of PPP, we will display the "fair value" of each currency against the Euro: EURCHF=1.929, EURAUD=1.288, EURCAD=1.228, EURUSD=1.062, and EURGBP=0.68.

Now, we will consider that the market knows the fair price for these pairs and we will write down the 130-simple-monthly average of each one. With the same order as in the previous point: 1.5016, 1.6509, 1.49, 1.2477, and 0.7353.

We will average the data from 1 and 2, and we will get what we call Simple Big Market Value: 1.7153 (Swiss), 1.46945 (Australian), 1.359 (Canadian), 1.1548 (US dollar), and 0.7076 (British).

Next, we will write their interest rates, using the excellent Global Rates, for instance. LIBOR 12 month of each currency will serve: 0.32%, 4.88%, 1.71%, 0.99%, and 1.79% for the pound. Then, we will extrapolate next-year inflation with the help of Index Mundi. How? We add last year's differential to the present inflation. Step by step. We will search "inflation Switzerland" within Index Mundi, and we will get this page. We can see that in 2010 inflation was -0.5% and in 2011: 0.7%. Then we will estimate that in 2012, inflation is going to be 0.7+(0.7-(-0.5))=1.9%. We do the same thing with each currency and get: 1.9%, 4%, 2.9%, 2.7%, and 4.4%.

We calculate the difference between the figures in point 4 and point 5 (RIR, real interest rates) and obtain: -1.36% (Swissy), 2.5% , -1.03% , -1.97% , and -2.82% (pound). Somehow, investors prefer to earn money through interest rates avoiding inflation.

We take these values into account by calculating (1+RIR)^10: 87.19%, 128.01%, 90.19%, 81.93%, and 75.09%, respectively. We call it correction. We are close, now.

Under no other circumstances, the usual weight should be 20% each currency (100%/5). However, we have seen some other factors. Let´s take them into account with the following formula: 20%*(ACTUAL PRICE/SBM)*CORRECTION, and we get: 12.19% (Swissy), 23.33% , 18.5% , 20.24% , and 18.82% (pound). (Note: we have used EURCHF= 1.2036, EURAUD=1.3445, EURCAD=1.3928, EURUSD=1.415, and EURGBP=0.8784.)

The problem is that the corrected weights don´t correspond with 100% but 93.08% in this case. Instead of 93.08% we want to show the weights for 100%, so linearly we just have to divide each weight by 93.08% and we finally get: CHF>13.1%, AUD>25.06%, CAD>19.87%, USD>21.75%, and GBP>20.22%.

The above follows a very practical approach. We just have to check out the portfolio every 6 months. In the next article, we will improve the results shown above.

Disclosure: I am long FXA, FXC, FXB, FXF, FXE.